The accuracy of physicians' perceptions of patients' suffering: findings from two teaching hospitals

How accurately physicians perceive patient suffering remains unclear. The authors sought to quantitatively compare physicians' estimates of their patients' suffering with the patients' ratings of their own suffering, using a paired survey.

Six major domains of suffering (DOSs) were derived from narrative descriptions of suffering by physicians and patients in a preliminary multicenter pilot study. From September 2005 through July 2006 at two teaching hospitals in Washington, DC, and Bethesda, Maryland, before a clinical encounter between a patient and physician, patients rated the impact of each of these DOSs on their overall suffering. After the same encounter, physicians rated the same DOSs according to their perception of suffering experienced by that patient. Patient responses were compared with physician responses using the Wilcoxon signed ranks and Spearman correlation tests.

Two hundred twenty-seven adult patients and their treating physicians completed the survey. Cooperation rates among patients and physicians were 94% and 97%, respectively. For two of the six DOSs (pain and physically nonpainful symptoms), there was no significant difference between the physicians' estimates of suffering and the patients' ratings of the DOS. For the remaining four DOSs (communication, emotional factors, loss, and systems factors), there was significant disagreement between the physicians' estimates and the actual suffering of the patients (P < .01). When all six DOSs were combined to ascertain how well perceptions of overall suffering correlated, significant discordance was also observed between physicians' perceptions and patients' descriptions (P < .001).

This study suggests that the physicians who participated might need more training in the recognition of patient suffering. Because these physicians were trained in medical schools across the country, the deficiencies noted here may not be limited to only the physicians in this study. More studies of physicians' ability to detect and manage suffering are needed, especially at nonteaching hospitals.